Followup Report on OD PMT Gain-Setting

K. Scholberg

Duke University

March 25, 2006

March 06 Dark Surveys

To check whether the tubes installed in the tank have gain-voltage relationships similar to those determined in the dark box last summer, I did two scope surveys during available dark test time on March 22, 2006. I had about 2.5 hours to work, and I looked at as many channels as possible during that time.

1. The first check was a survey of hut 2 tubes in the tank. I looked at the same tubes as shown in Figure 1 of the previous report, but had time for only 6 paddle cards' worth (total of 50 new tubes, where ``new'' means post-reconstruction tubes with a clear spe peak.)

2. The second check was a survey of 9 freshly-installed hut 1 tubes. For these I used the one cable connected to hut 1, crate 2, card 20, channel 9 (which I believe has no zener), and moved it from connector to connector at the cable hole baffle. I determined the cable-to-tube mappings from the spreadsheet sent by Bill.¹

I used channel 1 of a scope that I am pretty sure is the same one Katie and I used in the summer, a Tektronix 2465 belonging to UCI, which has a handy $\Delta V$ feature that allows relatively accurate pulse height measurements. Due to lack of time, I did not let the tubes settle for at least 5 minutes, as was done in the dark box last summer. (However in the summer we observed relatively few tubes changing at all after turn-on, so this shouldn't matter too much.)

Hut 2 Survey Results

The pulse height distribution of new tubes in hut 2 is shown in Figure 1. The pulse heights were quite different from those measured in the summer, with a mean value of 80 mV rather than 25 mV.

Figure 1: Distribution of spe pulse heights for new tubes in hut 2, crate 2, measured March 22, 2006.

I then noticed an ``interesting'' label on the side of the scope: see Figure 2.

Figure 2: Left: UCI oscilloscope, and the label on the side. Photos: Piotr Mijakowski

Figure 3 shows the July values (probably measured on channel 2) plotted versus the March ones (channel 1.) The values are entirely proportional, with a scale factor closer to 3 than 2 (as reported on the scope label.) It is very likely that we used channel 2 of this scope last summer, which accounts for the discrepancy. ²

Figure 3: Spe pulse heights for new tubes in hut 2, crate 2, measured in July 2005, probably on channel 2 of the UCI scope, compared to measurements made on channel 1 in March 2006. Points in red are the first two measurements made in July (see footnote 2 above.)

Since recommended voltages for new tubes were determined for a target pulse height of 25 mV, these same voltages should yield pulse heights of approximately 80 mV.

I also tried a different scope, a Tektronix TAS 425, which was harder to use due to lack of a $\Delta V$ feature: measured pulse heights are less accurate. I had to terminate this one at the connector in 50 ohms, since it does not have an internal 50 ohm termination feature. Results are shown in Figure 4. This one also showed a good correlation with channel 1 of the UCI scope, although there is again a scale factor (the slope is not unity.) Perhaps one or both scopes have their calibrations off, but in any case either would work fine for relative measurements.

Figure 4: Spe pulse heights for new tubes in hut 2, crate 2, measured on a TAS 425 scope, versus measurements on channel 1 of the UCI scope.

Results of Survey of Newly-Installed Tubes

For each of the nine tubes I measured pulse height on channel 1 of the UCI scope for a set of voltages ranging from around 1500 to 2100 V. I was running out of time and working quickly, but pulse height values should be good to about $\pm$5 mV. A few examples, for which the old measurements are scaled by 80./25., are shown in Figure 5.

Figure: For the first four tubes (indicated by different colors), the March 06 voltage-pulse height curve is shown as a solid line with circles; the rescaled July 05 measurements are shown as a dashed line with triangles.

Then, I calculated new target voltages for 80 mV pulse height. The results are shown in Table 1.

Table 1: Tubes with voltages recommended based on the dark box test, and new voltages based on March 06 measurements (note that quality of the former measurement is probably better.)

Paddle Card	Tube	July 05 recommended voltage (V)	80 mV voltage from Mar 06 measurement
1.2.04.07	SA3542	2041	2062
1.2.01.06	SA3946	2014	2051
1.1.03.03	SA4121	1800	1746
1.1.05.12	SA4143	1867	1894
1.1.08.09	SA4195	1998	2037
1.1.16.12	SA4199	1793	1783
1.1.12.10	SA4209	1777	1839
1.2.18.12	SA4244	1811	1920
1.1.20.12	SA4245	1781	1803

In all cases the new target voltages are within about 60 V of the old. There is one exception, SA4244. for which the difference is 109 V. I don't know if this is measurement error (I did not have a chance to check it), or if there's a big cable length difference in this case, or if the tube's characteristic changed, or if this happens to be one of the few tubes that needs to settle after being turned on.

Conclusion

After making the correction for scope calibration, gain-voltage relations are generally consistent with those from last summer. The one anomaly out of nine is slightly disturbing. It might be a good idea to resurvey in the tank before doing the final zener solder party, to refine the recommended voltages.

The lesson learned here (besides ``look carefully for labels on equipment, and then read them'') is to be aware that spe pulse height results can be scope-dependent. A particular scope may be used for reliable relative gain-setting, but needs to be calibrated relative to known levels.